Tamper Detection and Localization for Categorical Data Using Fragile Watermarks for Multi Owner Databases

摘要

With the development of the Internet and the information technology, databases can be easily manipulated and tampered. This creates a pressing need for integrity protection for databases. There are several existing watermarking techniques developed to detect the data tampering in single owner databases. However, if there are multiple owners of the data and if any of the owners is malicious or dishonest, he or she can tamper the data and re-embed the watermark which can't be detected. In this paper, we propose to extend the fragile watermarking algorithm proposed in the paper [1] to multi owner databases. In our algorithm, we divide the secret key into multiple shares using Shamir's secret sharing algorithm. All the tuples in a database are securely divided into groups using the master key. Within each group, the fragile watermark is embedded or verified with its corresponding owner's secret share. The design of this algorithm makes the groups more granular, thus, greatly increasing the probability of localizing and detecting the data tampering.